Abstract

Background: The incidence of gentamicin-associated acute kidney injury (AKI) as defined by the RIFLE criteria is unknown.

Aim and design: We performed a retrospective observational study to examine this and the predictive value of RIFLE stage on patient outcome in this setting.

Methods: We included all patients who were treated with gentamicin at our centre over a 1-month period. Data on 228 patients across all specialities were collected by manual searching of hospital notes and electronic pathology reporting systems. Information collected included baseline and peak serum creatinine results, gentamicin dose and serum levels, the presence of additional renal insults and the Stoke co-morbidity index.

Results: AKI occurred in 51 (24.4%) patients; 37 (17.7%) ‘Risk’, 9 (4.3%) ‘Injury’, 5 (2.4%) ‘Failure’. Independent predictors of gentamicin associated AKI were number of gentamicin levels >2 mg/l (OR 1.845, 95% CI 1.22 to 2.79) and higher baseline serum creatinine (OR 1.014, 95% CI 1.001–1.028). There was a greatly increased risk of in-hospital mortality in the AKI group as compared to those without AKI (45.1% vs. 19.1%, OR 3.48, 95% CI 1.8–6.9, P = 0.0004). Risk of in hospital mortality increased with each RIFLE stage (P < 0.0001).

Conclusions: This study shows that gentamicin-associated AKI remains a common and potentially serious clinical problem. There is a strong correlation between RIFLE class and in-hospital mortality.

Introduction

Gentamicin is an aminoglycoside antibiotic that has been in use for over 40 years. It is effective against a wide range of Gram-negative bacteria and is also active against Staphylococcus and Enterococcus, especially in synergy with β-lactams. Gentamicin remains a first-line antibiotic for many severe infections as in addition to clinical effectiveness the rates of resistance remain low and it is inexpensive. Moreover, the low risk of Clostridium difficile with gentamicin may further promote its use in preference to cephalosporins or quinolones.

The major drawbacks of the aminoglycosides are the requirement for monitoring plus the risk of nephrotoxicity and ototoxicity. There have been several factors previously reported as increasing the risk of nephrotoxicity; these include elevated trough gentamicin levels,1,2 plasma concentration-time area under the curve,3 duration of treatment2,4,5 concomitant vancomycin3 or frusemide use,5 volume depletion, elevated baseline serum creatinine4 and, possibly, increasing age.6 There are also reports of single doses of gentamicin causing nephrotoxicity.7 The reported rates of gentamicin nephrotoxicity vary widely from as low as 1.2%8 up to as high as 55%,9 although most studies report rates of between 8% and 26%.1,2,5,10–12 In part, some of this variation may be due to differences in defining nephrotoxicity and acute kidney injury (AKI) and there are no studies evaluating gentamicin toxicity since the widespread acceptance of the Acute Dialysis Qualitive Initiative Group's RIFLE criteria that provide a standardized definition of AKI.13 The RIFLE criteria have been shown to accurately predict prognosis in several clinical situations including intensive care, those undergoing cardiac or aortic surgery and in unselected hospital-wide admissions.14 However, gentamicin nephrotoxicity is one area in which the RIFLE scoring system has not been assessed and the current incidence of gentamicin associated AKI is not known when these criteria are employed.

There is no clear consensus as to the optimal model to monitor serum gentamicin levels, although once daily dosing may reduce the complexity of monitoring.15 Once daily dosing has been shown in some studies and one large meta-analysis to reduce but not eliminate the incidence of nephrotoxicty.3,8,10 Once daily dosing is possible due to the significant post-antibiotic effect of gentamicin, in that bacterial growth continues to be inhibited for 1–8 h after levels have dropped below the minimum inhibitory concentration (MIC).16 However, whichever monitoring method is used a certain degree of diligence is required as compared to alternative antibiotics, with one previous audit showing that clinical practice in this area is sometimes suboptimal.17

Therefore, we performed a retrospective observational study to evaluate the incidence of AKI in patients receiving once daily gentamicin as defined by RIFLE criteria. We sought to determine whether RIFLE score was predictive of outcome and identify any factors associated with the development of gentamicin nephrotoxicity.

Methods

Patients

All patients who were treated with gentamicin in our centre between 1 October and 1 November 2007 were identified from pharmacy dispensing records. All specialties were included except obstetrics and gynaecology who have separate hospital documentation that precluded complete data collection. In total, 245 patients were screened of which 228 were included in the study. The reasons for exclusion were notes unavailable (n = 14), patient on dialysis (n = 1) and two patients were subsequently found not to have received gentamicin.

Data collection

Data were collected by manual searching of hospital notes and electronic pathology reporting systems. At the time of the study, our hospital had in place antibiotic prescribing guidelines issued by the microbiology department governing gentamicin use. In an attempt to minimize errors regarding dose, gentamicin was prescribed using a dedicated chart that included simple instructions for dosing according to three weight bands (50–60, 61–95 and >95 kg), age and co-existing risk factors for nephrotoxicity. Schedule one (higher dose, 3.7–5.9 mg/kg) was recommended for patients aged <70 years with no co-existing risk factors (specified as ‘oliguria, cachexia, abnormal serum creatinine, hypotension, prescribed inotropes’); schedule two (2.5–3.9 mg/kg) was for patients 70 years or older or those with cautions. The on-call microbiologist was available to give dosing advice for patients weighing <50 kg. Trough gentamicin levels were checked 14–24 h after each dose and the dosing interval altered if levels were >1 mg/l. If the level was between 1 mg/l and 2 mg/l then dosing was changed to every 36 h. If the gentamicin level was >2 mg/l then no further doses were given and dosing strategy/antibiotic choice was discussed with on-call microbiologist. The complete antibiotic prescribing guidelines and prescription chart are included as appendices.

In addition to demographic details, information was collected regarding indication for gentamicin use, gentamicin dosing, gentamicin assay results, the accuracy of prescribing as per microbiology guidelines, the presence of additional renal insults and baseline and peak serum creatinine results to determine AKI as per RIFLE criteria.13 We used the standard definition of RIFLE criteria based on changes in serum creatinine, including the suggested addition of an absolute increase in creatinine of 26.4 μmol/l in the Risk category from the Acute Kidney Injury Network.18 We also used the RIFLE classification's suggestion of a period of 7 days over which to assess change in creatinine. This is summarised in Table 1. We did not include urine output criteria as urine output records were not available in all patients. Baseline creatinine was defined as the measurement immediately prior to receiving gentamicin in an attempt to exclude AKI occurring for reasons unrelated to gentamicin administration. Patients were excluded if serum creatinine was clearly rising prior to the administration of gentamicin. Additional potential renal insults were defined as hypotension (defined as a systolic pressure <100 mmHg), administration of intravenous contrast or the concurrent prescription of non-steroidal anti-inflammatory agents (NSAIDs), angiotensin converting inhibitors (ACEi), angiotensin receptor blockers (ARBs), diuretics or vancomycin. Peak serum creatinine was timed from the first gentamicin dose. All patients were categorised as having low (0), medium (1) or high (2) co-morbidity as per the Stoke scoring index19 and survival to discharge was recorded. In the absence of coding for the RIFLE classification of AKI, we also collected data regarding mortality and incidence of acute renal failure (ARF, as defined by ICD 10) throughout the entire hospital during the same study period.

Table 1

RIFLE criteria as modified by the Acute Kidney Injury Network (AKIN)

RIFLE class AKIN category Serum creatinine criteria Urine output criteria 

 
Risk An increase of ≥26.4 μmol/l (≥0.3 mg/dl) OR An increase of ≥150–200% (1.5- to 2-fold) from baseline <0.5 mg/kg/h for at least 6 h 
Injury An increase of ≥200–300% (2- to 3-fold) from baseline <0.5 mg/kg/h for at least 12 h 
Failure An increase of ≥300% (>3-fold) from baseline OR Serum creatinine ≥354 μmol/l (≥4.0 mg/dl) with an acute rise of at least 44 μmol/l (≥0.5 mg/dl) OR Initiation of RRT <0.3 mg/kg/h for at least 24 h OR Anuria for >12 h 
RIFLE class AKIN category Serum creatinine criteria Urine output criteria 

 
Risk An increase of ≥26.4 μmol/l (≥0.3 mg/dl) OR An increase of ≥150–200% (1.5- to 2-fold) from baseline <0.5 mg/kg/h for at least 6 h 
Injury An increase of ≥200–300% (2- to 3-fold) from baseline <0.5 mg/kg/h for at least 12 h 
Failure An increase of ≥300% (>3-fold) from baseline OR Serum creatinine ≥354 μmol/l (≥4.0 mg/dl) with an acute rise of at least 44 μmol/l (≥0.5 mg/dl) OR Initiation of RRT <0.3 mg/kg/h for at least 24 h OR Anuria for >12 h 

Baseline creatinine for the purpose of this study was taken to be the reading immediately prior to the administration of gentamicin. RRT: renal replacement therapy.

Statistical analysis

Results are expressed as mean ± SD unless otherwise stated. After demonstration of a normal distribution, two-tailed unpaired Student's t-tests were used to compare the distribution of quantitative variables and Fisher's exact test for categorical variables. Univariate analysis was performed to identify potential determinants of AKI and in-hospital mortality. Significant associations were then tested by binomial logistic regression analysis. All analysis was performed on an intention to treat basis. An alpha error at P < 0.05 was judged to be significant.

Results

Study population

A total of 136 (60%) patients were male and the mean age was 69.9 ± 17 years. Only 12% of patients were in intensive care at the time of AKI. The mean number of gentamicin doses administered was 3.4 ± 2 per patient (range 1–15) with a mean dose of 214 ± 85 mg (3.3 ± 1.1 mg/kg). Indications for gentamicin use are displayed in Figure 1. At baseline, the mean serum creatinine was 95.6 ± 28 μmol/l and mean peak serum creatinine was 124.4 ± 70 μmol/l (P < 0.0001 vs. baseline). Mean interval from first gentamicin dose to peak creatinine was 6.5 ± 9 days. Eighty-seven (38.2%) patients had a low co-morbidity score, 130 (57.0%) had a medium score and 11 (4.8%) had a high score. Age and co-morbidity score were correlated (Pearson r = 0.20, P = 0.003). The in-hospital mortality was 26.3% (60 patients).

Figure 1.

Frequency chart of indication for gentamicin use. Acute abdo: acute abdomen; CAP: community-acquired pneumonia; UTI/pyelo: complicated urinary tract infection/pyelonephritis; COPD: infective exacerbation of chronic obstructive pulmonary disease; HAP: hospital acquired pneumonia, neut sepsis: neutropaenic sepsis, other: an indication outside microbiology guidelines, not clear: indication unclear from hospital case notes review.

Figure 1.

Frequency chart of indication for gentamicin use. Acute abdo: acute abdomen; CAP: community-acquired pneumonia; UTI/pyelo: complicated urinary tract infection/pyelonephritis; COPD: infective exacerbation of chronic obstructive pulmonary disease; HAP: hospital acquired pneumonia, neut sepsis: neutropaenic sepsis, other: an indication outside microbiology guidelines, not clear: indication unclear from hospital case notes review.

AKI and gentamicin prescribing

Nineteen (8.3%) patients only had serum creatinine measured at presentation that precluded determination of AKI. Of the remaining 209 patients, 51 (24.4%) developed AKI as per the RIFLE criteria. Thirty-seven (17.7%) patients fell into the ‘Risk’ category, nine (4.3%) were in the ‘Injury’ category and five (2.4%) were in the ‘Failure’ group. No patients required renal replacement therapy.

Table 2 summarizes the comparison of patients with and without AKI. Notably, patients who developed AKI were older, had a higher mean baseline serum creatinine and a trend towards higher co-morbidity scores. This may explain why this group were more likely to have been dosed as per schedule 2 (88.6% vs. 61.6% in non-AKI patients, OR 4.76, 95% CI 1.82–14.4, P = 0.0007) and therefore had a lower mean gentamicin dose. However, there was no difference in the proportion of those developing AKI when patients were divided into those who received a dose of ⩽3 mg/kg and those who received >3 mg/kg (20% vs. 17.9%, respectively, OR 1.1, 95% CI 0.5–2.6, P = 0.71). Furthermore, we did not observe any differences between the two groups in duration of gentamicin course (number of doses).

Table 2

Comparison of patients with and without AKI

 AKI group Non-AKI group P-value 

 
Age 74.5 ± 13 68.5 ± 18 0.031 
Max. gentamicin trough level (mg/l) 2.4 ± 1.7 1.4 ± 1.7 0.0017 
No. of levels >2 mg/l per patient 1.1 ± 1.3 0.38 ± 0.77 <0.0001 
No. of patients with at least one level >2 mg/l 24 (47%) 34 (22%) 0.001 
No. of levels >5 mg/l per patient 0.1 ± 0.3 0.02 ± 0.2 0.049 
No. of patients with at least one level >5 mg/l 4 (8%) 3 (2%) 0.36 
Baseline serum creatinine (μmol/l) 105.9 ± 34 92.1 ± 26 0.0026 
Peak serum creatinine (μmol/l) 199.8 ± 102 99.8 ± 33 <0.0001 
Percent rise in serum creatinine 94.2 ± 103 9.5 ± 11 <0.0001 
Body weight (kg) 68.2 ± 18 69.1 ± 16 0.75 
Dose (mg) 192 ± 65 221 ± 90 0.038 
Dose per weight (mg/kg) 3.09 ± 1.0 3.3 ± 1.1 0.46 
No. of doses per pt. 3.4 ± 2 3.6 ± 3 0.56 
No. of gentamicin levels per patient 4.2 ± 4 3.5 ± 3 0.92 
Percentage of patients with additional renal insults (%) 81.3 62.1 0.014 
Co-morbidity    
    0 29.4% 40.5%  
    1 60.8% 56.2% 0.053 
    2 9.8% 3.3%  
Percentage of in-hospital death 45.1% 19.1% 0.0004 
 AKI group Non-AKI group P-value 

 
Age 74.5 ± 13 68.5 ± 18 0.031 
Max. gentamicin trough level (mg/l) 2.4 ± 1.7 1.4 ± 1.7 0.0017 
No. of levels >2 mg/l per patient 1.1 ± 1.3 0.38 ± 0.77 <0.0001 
No. of patients with at least one level >2 mg/l 24 (47%) 34 (22%) 0.001 
No. of levels >5 mg/l per patient 0.1 ± 0.3 0.02 ± 0.2 0.049 
No. of patients with at least one level >5 mg/l 4 (8%) 3 (2%) 0.36 
Baseline serum creatinine (μmol/l) 105.9 ± 34 92.1 ± 26 0.0026 
Peak serum creatinine (μmol/l) 199.8 ± 102 99.8 ± 33 <0.0001 
Percent rise in serum creatinine 94.2 ± 103 9.5 ± 11 <0.0001 
Body weight (kg) 68.2 ± 18 69.1 ± 16 0.75 
Dose (mg) 192 ± 65 221 ± 90 0.038 
Dose per weight (mg/kg) 3.09 ± 1.0 3.3 ± 1.1 0.46 
No. of doses per pt. 3.4 ± 2 3.6 ± 3 0.56 
No. of gentamicin levels per patient 4.2 ± 4 3.5 ± 3 0.92 
Percentage of patients with additional renal insults (%) 81.3 62.1 0.014 
Co-morbidity    
    0 29.4% 40.5%  
    1 60.8% 56.2% 0.053 
    2 9.8% 3.3%  
Percentage of in-hospital death 45.1% 19.1% 0.0004 

The patients who developed AKI also had a higher number of trough gentamicin levels >2 mg/l and >5 mg/l per patient as compared to the non-AKI group. In addition, the maximum gentamicin trough level was significantly higher in the AKI group (2.4 ± 1.7 mg/l vs. 1.4 ± 1.7 mg/l, P = 0.0017). Of all patients who had gentamicin levels checked, 27 (15%) had more than one trough level of >2 mg/l. 14 (51.9%) of these 27 developed AKI.

A significantly higher proportion of patients with AKI had one or more additional renal insults as compared to the non-AKI group (81.3% vs. 66.7%, OR 2.65, 95% CI 1.19–5.86, P = 0.014). Overall, 166 (66%) patients received a total of 215 additional renal insults, the nature of which are summarized in Figure 2.

Figure 2.

Frequency of additional factors that may potentially contribute to the development of AKI. NSAIDs = non-steroidal anti-inflammatory agents; ACEi/ARB = angiotensin converting inhibitors or angiotensin receptor blockers.

Figure 2.

Frequency of additional factors that may potentially contribute to the development of AKI. NSAIDs = non-steroidal anti-inflammatory agents; ACEi/ARB = angiotensin converting inhibitors or angiotensin receptor blockers.

Logistic regression identified the number of gentamicin levels >2 mg/l (OR 1.845, 95% CI 1.22–2.79) and baseline serum creatinine (OR 1.014, 95% CI 1.001–1.028) as the only independent predictors of AKI (Nagelkerke r2 = 0.178).

Of those patients with AKI who survived, 21 (78%) had complete recovery of renal function at the point of discharge from hospital and six (22%) had a partial recovery. Of those with a partial recovery, mean serum creatinine was 172 ± 34 μmol/l and mean estimated GFR (MDRD) was 31 ± 10 ml/min. This resulted in four of these patients being classified as CKD stage 4, and two CKD stage 3, although we did not collect longer term follow-up data that may have revealed further recovery over time. In the patients with AKI who died, only five (21%) patients recovered renal function, one (4%) had partial recovery and 18 (75%) died without any improvement in renal function.

In view of their importance in the development of AKI, we identified factors that may predict high gentamicin trough levels. We found that those who had at least one trough level >2 mg/l were older (73.3 ± 15 years vs. 66.8 ± 18 years, P = 0.004) and had higher baseline serum creatinine values (105 ± 31 μmol/l vs. 87 ± 23 μmol/l, P < 0.0001). As seen for the overall AKI group, there was no influence of gentamicin dose.

Overall, gentamicin prescribing was not done well despite the widespread availability of guidelines and prescribing instructions. 46 (21.6%) patients were prescribed the incorrect schedule and 92 (42.2%) patients were classified as having the incorrect dose of gentamicin prescribed. However, this included 52 (23.9%) patients for whom there was a failure to record body weight. Fifteen of these patients had a gentamicin dose prescribed outside of the advised ranges, whilst for 37 patients it was not possible to determine whether the dose would have been correct if the weight had been recorded (i.e. scored as incorrect dose due to lack of weight only). If these patients are excluded, then 55 (30.4%) of patients were dosed incorrectly despite documentation of body weight. Conversely, monitoring of serum gentamicin levels was performed better. The mean number of serum gentamicin levels were 3.5 ± 3.5 per patient, slightly higher than the mean number of doses per patient and only 10 (4.4%) patients had missed gentamicin levels.

There was no association between incorrect gentamicin prescribing and AKI; the correct schedule was chosen in 76.6% of AKI patients as compared to 79.3% in the non-AKI group (OR 0.81, 95% CI 0.37–1.78, P = 0.67), whilst the correct dose was prescribed in 52% of AKI group as compared to 59.1% of the non-AKI group (OR 0.75, 95% CI 0.39–1.43, P = 0.41). 72.5% of those with AKI and 72.6% of those without AKI had body weight recorded.

In-hospital mortality

There was a greatly increased risk of in-hospital mortality in the AKI group as compared to those without AKI (45.1% vs. 19.1%, OR 3.48, 95% CI 1.8–6.9, P = 0.0004). The percentage of in-hospital mortality increased with each stage of the RIFLE criteria as summarized in Figure 3 (chi-squared test for trend P < 0.0001). Other significant associations with in-hospital death on univariate analysis were increasing age and co-morbidity. Mean age of patients who survived to discharge was 67.3 ± 18 years as compared to 77.0 ± 12 years in those who died (P = 0.0001). The proportion of patients who died who were classified as low, medium and high co-morbidity was 15.3, 78.0 and 6.7%, respectively, as compared to 44.8, 50.9 and 4.3% in the survivors (chi-squared test for trend P = 0.0002).

Figure 3.

Proportion of in-patient mortality stratified by RIFLE criteria. The chi-squared test for trend highly significant (P < 0.0001) and logistic regression analysis calculated an odds ratio of 2.62 for each increase in RIFLE score (95% CI 1.59–4.31).

Figure 3.

Proportion of in-patient mortality stratified by RIFLE criteria. The chi-squared test for trend highly significant (P < 0.0001) and logistic regression analysis calculated an odds ratio of 2.62 for each increase in RIFLE score (95% CI 1.59–4.31).

All of these factors were identified as independent predictors of in-hospital mortality on multivariate analysis; RIFLE criteria (OR 2.504 for each increase in RIFLE score, 95% CI 1.46–4.3), age (OR 1.05 for each year increase, 95% CI 1.02–1.1) and co-morbidity score (OR 2.377 for each increase in score, 95% CI 1.13–5.0, Nagelkerke r2 = 0.259).

Hospital-wide data

We also collected data regarding mortality and incidence of ARF as defined by ICD 10 coding throughout the entire hospital during the same study period, for which there were a total of 5880 admissions and 196 deaths (3.3%). Of these, 1826 admissions were to internal medicine of which there were 121 deaths (6.6%). Clinical coding identified 86 additional patients who were not included in the gentamicin study population who had ‘acute renal failure’ and of these 21 died (24.4%).

Discussion

This study confirms the utility of the RIFLE criteria in predicting outcomes of AKI associated with gentamicin use. Our reported incidence of gentamicin-associated AKI is similar to previously reported rates,1,2,5,10–12 but nephrotoxicity is a significant problem occurring in almost a quarter of patients prescribed gentamicin.

Our results show that the RIFLE criteria perform well in predicting outcome in gentamicin-associated AKI. Although the rates of AKI we observed are similar to other reported rates of gentamicin nephrotoxicity,1,2,5,10–12 it is clear that these rates are higher than in some studies, which in part may be due to differing definitions of nephrotoxicity.8 However, as even small rises in serum creatinine are associated with a poor outcome,20 gentamicin nephrotoxicity is clinically important as a common, potentially modifiable renal insult.

Although the exact intracellular mechanisms are not entirely clear, gentamicin nephrotoxicity occurs after uptake into proximal tubular cells and subsequent accumulation in lysosomes. Cellular uptake occurs by calcium dependent active transport that becomes saturated at levels attained with standard dosing.16 Prolonged trough levels >2 mg/l increase the time over which gentamicin uptake occurs, resulting in greater intracellular gentamicin concentrations and therefore greater toxicity.16 Certainly, we found that elevated serum trough levels (>2 mg/l) predicted the development of AKI in our study, a finding also reported elsewhere.1,2 The elevated peak serum gentamicin levels seen with once daily dosing do not appear to increase nephrotoxicity providing gentamicin clearance is preserved; the higher peaks cannot increase the rate of gentamicin uptake into the proximal tubules once this process is saturated and with one daily dose there is less time with high trough levels.3,8–10 However, patients with abnormal renal function have reduced gentamicin clearance and are more likely to have longer periods with elevated serum levels. Our results and those from other studies show that abnormal baseline renal function is also an important predictor of AKI.4,6 Prins et al.5 clearly demonstrate that for patients with normal renal function, once daily gentamicin dosing achieves satisfactory peak levels (at least six times that of the MIC) whilst avoiding high trough levels. However, when creatinine clearance is reduced below 50 ml/min and especially when <30 ml/min, even a lower starting dose (2–2.5 mg/kg) may result in a significant proportion of patients with 24-h trough levels of >2 mg/l, at the same time failing in some cases to achieve adequate peak levels.5

We also observed that increasing age and abnormal baseline creatinine were associated with a higher risk of gentamicin trough levels >2 mg/l (although there are several studies demonstrating, as we have that age per se is not an independent risk factor for gentamicin nephrotoxicity).2,4 This pharmacokinetic profile suggests that once daily gentamicin should be avoided as much as possible in patients with impaired creatinine clearance or used only under specialist supervision, especially when there are additional risk factors for AKI. This is particularly pertinent in view of observations that show gentamicin prescribing and monitoring is sometimes suboptimal in routine clinical practice.17

Some previous studies have identified the duration of gentamicin course to be associated with AKI.2,4,5,21 Our results did not show this, which may be due to the relatively low mean number of doses per patient. However, this finding should be interpreted with caution as we have used number of doses as surrogate for duration of gentamicin use. One study has found that the risk of nephrotoxicity was only increased if the length of treatment exceeded 11 days.2 Although not all studies have found a significant influence on duration of therapy and risk of nephrotoxicity,22 it seems prudent to limit the duration of gentamicin courses as much as possible, ideally to <5–7 days to avoid potential drug accumulation, especially in those whose creatinine clearance is in flux.4

Interestingly, we found that there was no association between gentamicin dose and the development of AKI, which has also been reported by others.22,23 Indeed, there are no studies of which we are aware that do report a direct effect of gentamicin dose. Our results may in part reflect the fact that the doses prescribed were fairly conservative, with a mean dose of 214 ± 85 mg (3.3 ± 1.1 mg/kg). However, what may be more important is whether the dose is correctly individualized to the patient, particularly in terms of creatinine clearance in light of the association we and others report between abnormal renal function, elevated trough levels and AKI.1,2

Other studies have found that the presence of additional renal insults including concomitant vancomycin3 or frusemide use,5 increase the risk of nephrotoxicity. We also observed this trend, although the presence of additional renal insults was not an independent predictor of AKI on multivariate analysis. It is likely that some of the cases of AKI identified in our study were multi-factorial, due to several insults of which gentamicin was only one. Although this could be cited as a potential weakness of our study, we take the view that our results are representative of day-to-day clinical practice as gentamicin is often used in those with severe sepsis with all of its attendant risk factors for AKI. However, if gentamicin is used in those patients who are stable with normal renal function and low risk of additional insults, then the incidence of nephrotoxicity is likely to be <24%.

Our study did suffer from some limitations. In particular, it was a retrospective study and therefore suffers from the limitations associated with this, in particular that we cannot prove causation between gentamicin and the episodes of AKI that we observed. Furthermore, it was not possible to collect data quantifying severity of the acute illness. As mentioned earlier, we used ICD 10 coding to obtain data with regard to hospital wide ARF that makes comparison with AKI as defined by the RIFLE criteria difficult. In addition, we were unable to accurately collect information regarding urine output and our RIFLE classification was solely based on changes in serum creatinine. Finally, we used number of doses as a surrogate for duration of gentamicin treatment.

Conclusions

In conclusion, we have demonstrated that gentamicin-associated AKI as defined by the RIFLE criteria remains a common and potentially modifiable clinical problem, particularly when renal function is abnormal. This is highly relevant as the severity of AKI is a strong predictor of mortality.

Conflict of interest: None declared.

References

1
Matzke
GR
Lucarotti
RL
Shapiro
HS
Controlled comparison of gentamicin and tobramycin nephrotoxicity
Am J Nephrol
 , 
1983
, vol. 
3
 (pg. 
11
-
7
)
2
Raveh
D
Kopyt
M
Hite
Y
Rudensky
B
Sonnenblick
M
Yinnon
AM
Risk factors for nephrotoxicity in elderly patients receiving once-daily aminoglycosides
Qjm
 , 
2002
, vol. 
95
 (pg. 
291
-
7
)
3
Rybak
MJ
Abate
BJ
Kang
SL
Ruffing
MJ
Lerner
SA
Drusano
GL
Prospective evaluation of the effect of an aminoglycoside dosing regimen on rates of observed nephrotoxicity and ototoxicity
Antimicrob Agents Chemother
 , 
1999
, vol. 
43
 (pg. 
1549
-
55
)
4
Paterson
DL
Robson
JM
Wagener
MM
Risk factors for toxicity in elderly patients given aminoglycosides once daily
J Gen Intern Med
 , 
1998
, vol. 
13
 (pg. 
735
-
9
)
5
Prins
JM
Weverling
GJ
de Blok
K
van Ketel
RJ
Speelman
P
Validation and nephrotoxicity of a simplified once-daily aminoglycoside dosing schedule and guidelines for monitoring therapy
Antimicrob Agents Chemother
 , 
1996
, vol. 
40
 (pg. 
2494
-
9
)
6
Fujita
K
Sayama
T
Abe
S
Murayama
T
Tashiro
H
Age-dependent aminoglycoside nephrotoxicity
J Urol
 , 
1985
, vol. 
134
 (pg. 
596
-
7
)
7
Weir
BA
Mazumdar
DC
Aminoglycoside nephrotoxicity following single-dose cystoscopy prophylaxis
Ann Pharmacother
 , 
1994
, vol. 
28
 (pg. 
199
-
201
)
8
Nicolau
DP
Freeman
CD
Belliveau
PP
Nightingale
CH
Ross
JW
Quintiliani
R
Experience with a once-daily aminoglycoside program administered to 2,184 adult patients
Antimicrob Agents Chemother
 , 
1995
, vol. 
39
 (pg. 
650
-
5
)
9
Kumin
GD
Clinical nephrotoxicity of tobramycin and gentamicin. A prospective study
JAMA
 , 
1980
, vol. 
244
 (pg. 
1808
-
10
)
10
Barza
M
Ioannidis
JP
Cappelleri
JC
Lau
J
Single or multiple daily doses of aminoglycosides: a meta-analysis
BMJ
 , 
1996
, vol. 
312
 (pg. 
338
-
45
)
11
Schentag
JJ
Plaut
ME
Cerra
FB
Comparative nephrotoxicity of gentamicin and tobramycin: pharmacokinetic and clinical studies in 201 patients
Antimicrob Agents Chemother
 , 
1981
, vol. 
19
 (pg. 
859
-
66
)
12
Smith
CR
Lipsky
JJ
Laskin
OL
Hellmann
DB
Mellits
ED
Longstreth
J
, et al.  . 
Double-blind comparison of the nephrotoxicity and auditory toxicity of gentamicin and tobramycin
N Engl J Med
 , 
1980
, vol. 
302
 (pg. 
1106
-
9
)
13
Bellomo
R
Ronco
C
Kellum
JA
Mehta
RL
Palevsky
P
Acute renal failure – definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group
Crit Care
 , 
2004
, vol. 
8
 (pg. 
R204
-
12
)
14
Ricci
Z
Cruz
D
Ronco
C
The RIFLE criteria and mortality in acute kidney injury: A systematic review
Kidney Int
 , 
2008
, vol. 
73
 (pg. 
538
-
46
)
15
Cooke
RP
Grace
RJ
Gover
PA
Audit of once-daily dosing gentamicin therapy in neutropenic fever
Int J Clin Pract
 , 
1997
, vol. 
51
 (pg. 
229
-
31
)
16
Begg
EJ
Barclay
ML
Kirkpatrick
CM
The therapeutic monitoring of antimicrobial agents
Br J Clin Pharmacol
 , 
2001
, vol. 
52
 
Suppl. 1
(pg. 
35S
-
43S
)
17
Shrimpton
SB
Milmoe
M
Wilson
AP
Felmingham
D
Drayan
S
Barrass
C
, et al.  . 
Audit of prescription and assay of aminoglycosides in a UK teaching hospital
J Antimicrob Chemother
 , 
1993
, vol. 
31
 (pg. 
599
-
606
)
18
Mehta
RL
Kellum
JA
Shah
SV
Molitoris
BA
Ronco
C
Warnock
DG
, et al.  . 
Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury
Crit Care
 , 
2007
, vol. 
11
 pg. 
R31
 
19
Davies
SJ
Phillips
L
Naish
PF
Russell
GI
Quantifying comorbidity in peritoneal dialysis patients and its relationship to other predictors of survival
Nephrol Dial Transplant
 , 
2002
, vol. 
17
 (pg. 
1085
-
92
)
20
Chertow
GM
Burdick
E
Honour
M
Bonventre
JV
Bates
DW
Acute kidney injury, mortality, length of stay, and costs in hospitalized patients
J Am Soc Nephrol
 , 
2005
, vol. 
16
 (pg. 
3365
-
70
)
21
Eltahawy
AT
Bahnassy
AA
Aminoglycoside prescription, therapeutic monitoring and nephrotoxicity at a university hospital in Saudi Arabia
J Chemother
 , 
1996
, vol. 
8
 (pg. 
278
-
83
)
22
Moore
RD
Smith
CR
Lipsky
JJ
Mellits
ED
Lietman
PS
Risk factors for nephrotoxicity in patients treated with aminoglycosides
Ann Intern Med
 , 
1984
, vol. 
100
 (pg. 
352
-
7
)
23
Bartal
C
Danon
A
Schlaeffer
F
Reisenberg
K
Alkan
M
Smoliakov
R
, et al.  . 
Pharmacokinetic dosing of aminoglycosides: a controlled trial
Am J Med
 , 
2003
, vol. 
114
 (pg. 
194
-
8
)